Informational economic transmission is important (even after controlling for countries' fundamental real and financial linkages), where the informational interdependence emerges from anomalous co-movements in agents' beliefs about the economic performance between countries. We propose a novel measure of informational interdependence, the correlation (between countries) of analysts' GDP one-year-forecast errors, which is based on a stylized learning model, where informational links arise when agents over-weight the information precision about common-factors (due to learning costs and category-learning). We find that the informational interdependence changes during crises and between economies from different country-categories. Furthermore, we show substantial higher-order spillover amplifications of economic shocks.
We explore how comovement in stock returns can be explained by firm-to-firm interdependence through informational links. Due to limited attention capacity, agents' learning process is biased towards common information, which generates correlated beliefs about stock returns. This will generate interdependence across stock returns, and stock market comovement at an aggregate level. We use a novel measure for informational linkages between firms based on analysts forecast errors, estimate the relevance of this channel with a Spatial Two-Stage Least Square (S2SLS) estimator and find that the informational channel explains ``idiosyncratic'' returns, has a lower intensity in periods of higher uncertainty, stands and is amplified when taking into account category-learning effects. We also study the propagation of climate events and simulated shocks in the stock market based on the estimated informational linkages and find quantitatively important indirect effects.